U.S. patent application number 16/092689 was filed with the patent office on 2019-05-02 for real-time, parallel x-ray tomosynthesis.
The applicant listed for this patent is The Regents of the University of California. Invention is credited to Scott S. Hsieh.
Application Number | 20190126070 16/092689 |
Document ID | / |
Family ID | 60041899 |
Filed Date | 2019-05-02 |
![](/patent/app/20190126070/US20190126070A1-20190502-D00000.png)
![](/patent/app/20190126070/US20190126070A1-20190502-D00001.png)
![](/patent/app/20190126070/US20190126070A1-20190502-D00002.png)
![](/patent/app/20190126070/US20190126070A1-20190502-D00003.png)
![](/patent/app/20190126070/US20190126070A1-20190502-D00004.png)
![](/patent/app/20190126070/US20190126070A1-20190502-D00005.png)
![](/patent/app/20190126070/US20190126070A1-20190502-D00006.png)
![](/patent/app/20190126070/US20190126070A1-20190502-D00007.png)
![](/patent/app/20190126070/US20190126070A1-20190502-D00008.png)
![](/patent/app/20190126070/US20190126070A1-20190502-D00009.png)
![](/patent/app/20190126070/US20190126070A1-20190502-D00010.png)
View All Diagrams
United States Patent
Application |
20190126070 |
Kind Code |
A1 |
Hsieh; Scott S. |
May 2, 2019 |
Real-Time, Parallel X-Ray Tomosynthesis
Abstract
A device for performing tomosynthesis in real time is described.
Multiple imaging sources (such as x-ray sources) may be energized
in parallel and collimated towards a field of view. Objects within
the field of view cast shadows onto one or more detectors. An
imaging system may read the one or more detectors and acquire
multiple views corresponding to the multiple imaging sources to
produce a reconstructed image of an object of interest. From this
reconstructed image, a target of the radiation therapy can be
located, and the delivery of the radiation can be adjusted, as
needed. The approach provides a real-time tomosynthesis design that
can produce enhanced contrast for guidance of, for example, lung
tumor treatment. Higher frame rates can be achieved to better
compensate for changes in the position of the target during
radiation therapy due to, for example, respiratory or cardiac
motion.
Inventors: |
Hsieh; Scott S.; (Anaheim,
CA) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
The Regents of the University of California |
Oakland |
CA |
US |
|
|
Family ID: |
60041899 |
Appl. No.: |
16/092689 |
Filed: |
April 10, 2017 |
PCT Filed: |
April 10, 2017 |
PCT NO: |
PCT/US2017/026802 |
371 Date: |
October 10, 2018 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
62320788 |
Apr 11, 2016 |
|
|
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
A61B 6/06 20130101; A61B
6/4429 20130101; G01T 1/161 20130101; A61B 8/13 20130101; A61B
6/025 20130101; A61B 6/461 20130101; A61N 5/1067 20130101; A61B
6/4085 20130101; G06T 2211/424 20130101; A61N 5/1049 20130101; A61B
6/12 20130101; G01T 1/17 20130101; A61B 6/54 20130101; A61B 6/0487
20200801; A61B 6/4014 20130101; A61N 2005/1061 20130101; G01T 7/00
20130101; A61N 5/1068 20130101; G06T 11/006 20130101; A61B 6/00
20130101; A61N 5/1037 20130101; A61B 6/0407 20130101; A61B 6/486
20130101; A61B 6/4007 20130101; A61N 5/1045 20130101 |
International
Class: |
A61N 5/10 20060101
A61N005/10; A61B 6/06 20060101 A61B006/06; A61B 6/02 20060101
A61B006/02; A61B 6/00 20060101 A61B006/00; G06T 11/00 20060101
G06T011/00; G01T 1/161 20060101 G01T001/161; G01T 1/17 20060101
G01T001/17; G01T 7/00 20060101 G01T007/00 |
Claims
1. An apparatus for tomosynthesis imaging, comprising: a plurality
of imaging sources configured to, when energized, emit radiation
beams intersecting at a field of view and projecting images onto an
imager having a detection area; a collimator configured to
collimate the radiation emanating from each imaging source such
that the radiation from any single imaging source encompasses an
area substantially smaller than the detection area; and a
controller configured to: energize the plurality of imaging
sources; acquire readouts from the imager at a frame rate, each
readout including images corresponding to each energized imaging
source; and combine the acquired images to generate a reconstructed
image.
2. The apparatus of claim 1, wherein a reconstructed image is
formed for each detector readout.
3. The apparatus of claim 1, wherein the imaging sources are
aligned such that projections thereof are non-overlapping in the
detection area of the imager.
4. The apparatus of claim 1, wherein the controller is further
configured to estimate a location of a biological target of
interest based on the reconstructed image.
5. The apparatus of claim 4, wherein the controller is further
configured to determine an adjustment to delivery of high-energy
treatment radiation based on the estimated location.
6. The apparatus of claim 1, further including an imager coupled to
the controller and configured to communicate detected images
thereto.
7. The apparatus of claim 1, wherein the reconstructed image is
generated using shift-and-add.
8. The apparatus of claim 1, wherein the reconstructed image is
generated using filtered backprojection.
9. The apparatus of claim 1, wherein the reconstructed image is
generated using iterative reconstruction.
10. The apparatus of claim 1, wherein the acquired images are
combined so as to focus on a targeted feature and de-focus
surrounding features.
11. The apparatus of claim 1, wherein the imaging sources are
energized substantially simultaneously.
12. The apparatus of claim 1, wherein the imaging sources are
energized in rapid succession.
13. The apparatus of claim 1, wherein a first subset of the
plurality of imaging sources is energized for a first frame, and a
second subset of the plurality of imaging sources is energized for
a second frame that is subsequent to the first frame.
14. The apparatus of claim 13, wherein the plurality of imaging
sources is partitioned such that each imaging source is included in
only one of the first and second subsets.
15. The apparatus of claim 1, wherein each of the imaging sources
includes a kilovolt x-ray tube.
16. The apparatus of claim 1, further comprising a treatment
radiation source.
17. The apparatus of claim 16, wherein the treatment radiation
source is a linear accelerator or high-energy radioisotope
source.
18. The apparatus of claim 16, wherein the imaging sources are
positioned around the source of treatment radiation.
19. The apparatus of claim 18, wherein the imager is configured to
detect both kilovoltage photons and megavoltage photons.
20. The apparatus of claim 1, wherein the imager includes multiple
detectors.
21. The apparatus of claim 20, wherein each of the multiple
detectors receives projections from a subset of the plurality of
the imaging sources.
22. The apparatus of claim 1, wherein the apparatus is configured
to be integrated with a radiation therapy system having a
high-energy treatment radiation source such that the plurality of
imaging sources is arranged around the treatment radiation
source.
23. A method of guiding radiation therapy, comprising: energizing
multiple imaging sources such that radiation beams emitted by the
imaging sources intersect at a field of view and project shadow
images onto an imager; acquiring a readout image from the imager,
wherein the readout image includes shadow images corresponding with
the multiple imaging sources; generating a reconstructed image
using the shadow images of the multiple imaging sources; estimating
a location of a target of interest; and adjusting delivery of
high-energy treatment radiation to the target of interest based on
the estimated location.
24. The method of claim 23, wherein the multiple imaging sources
are arranged about a treatment source, wherein the treatment source
is used to deliver the high-energy treatment radiation.
25. The method of claim 23, wherein the target of interest is a
tumor.
26. The method of claim 23, wherein the adjusting delivery includes
gating a high-energy radiation source used to deliver the
high-energy treatment radiation.
27. The method of claim 23, wherein the adjusting delivery includes
re-aiming a high-energy radiation source used to deliver the
high-energy treatment radiation.
28. The method of claim 23, wherein the imager is configured to
detect both kilovoltage photons and megavoltage photons.
29. The method of claim 23, further comprising displaying the
reconstructed image to a user.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application is based on, claims priority to, and
incorporates herein by reference in its entirety U.S. Provisional
Application Ser. No. 62/320,788, filed Apr. 11, 2016, and entitled,
"Real-Time, Parallel X-Ray Tomosynthesis." The references cited in
the above provisional patent application are also hereby
incorporated by reference.
FIELD OF THE INVENTION
[0002] This invention relates generally to radiological imaging
with enhanced frame rates, and in particular, to tomosynthesis
well-suited for compensating for movements of patients during, for
example, radiation therapy.
BACKGROUND
[0003] Patient motion has been a recurring problem in medical
imaging. While computerized tomography (CT) and magnetic resonance
imaging (MRI) offer unparalleled accuracy and resolution, data
acquisition can require hundreds of milliseconds and may not
adequately resolve changes in positions of targeted features
resulting from, for example, respiratory or cardiac motion. With
radiation therapy for lung cancer, for example, lung tumor motion
is not adequately compensated for using only external surrogates
such as respiratory gating, and the speed and path of motion may
change between subsequent treatment fractions. In current lung
stereotactic ablative radiation therapy (SABR) treatments, an
additional internal target volume (ITV) margin is sometimes added
to account for the uncertainty in the current position of a target.
However, this increases the volume of tissue that receives
radiation, and as a result, surrounding tissue that is "normal"
(i.e., undiseased or otherwise untargeted) is more likely to
receive lethal radiation. SABR treatments of central tumors in
particular are associated with high levels of toxicity and tissue
necrosis, even with gentler fractionation schedules.
[0004] Fluoroscopy has been a tool of choice for monitoring
treatments or interventions because of its fast imaging times.
However, the contrast obtained in fluoroscopy is insufficient for
certain clinical applications. Tomosynthesis is sometimes used in
an effort to improve conspicuity (or contrast) relative to
fluoroscopy. In tomosynthesis, multiple x-ray images taken at
different angles are reconstructed to show a region of interest.
However, existing implementations of tomosynthesis require the
serial acquisition of several views. For a typical flat panel
imager acquiring data at 30 frames per second (fps), acquisition of
a tomosynthetic image could require half a second or more, which
does not provide the speed necessary for real-time feedback on
changes in position.
[0005] It would therefore be useful to have a tomosynthesis system
that can provide more rapid feedback to allow for real-time
tracking of targets, allowing for adjustments to radiation delivery
that are better able to compensate for patient motion during
treatment.
SUMMARY OF THE PRESENT DISCLOSURE
[0006] In exemplary implementations, the disclosed systems and
methods use multiple imaging sources to image a (potentially
moving) target with enhanced speed and contrast. The imaging
sources may simultaneously illuminate a field of view (that
includes a target of interest), and one or more detectors can be
used to receive the beams. In certain configurations, if the field
of view is sufficiently small, a single detector (e.g., a flat
panel detector) could be used by, for example, assigning different
sectors on the detector to different sources. The images can then
be combined to form a reconstructed image of the target. Such a
real-time tomosynthesis system can provide guidance for radiation
therapy for applications in which fluoroscopy or ultrasound provide
insufficient contrast, and/or diagnostic scanners such as MRI or CT
are too slow or unavailable in the clinical workflow. This
tomosynthesis approach could resolve, for example, cardiac or
respiratory motion during radiation treatments, such as treatments
for lung cancer.
[0007] An exemplary device for parallel tomosynthesis, in certain
configurations, includes of a plurality of small x-ray sources that
are tightly collimated down to a small region of interest in a
patient. These x-ray sources may be energized substantially
simultaneously or in rapid succession (e.g., as rapid as the
controller may instruct the imaging sources and the imaging sources
may be energized in sequentially). After traveling through the
patient, the images can be detected simultaneously in a single
readout of an x-ray detector, and used to produce a reconstructed
image. The reconstructed image can be used to make adjustments to
the delivery of high-energy treatment radiation to a moving
target.
[0008] An exemplary method for image guidance during high-energy
radiation therapy, in certain implementations, may use a plurality
of x-ray sources that are energized substantially simultaneously or
in rapid succession to image a region of interest within a patient.
A single detector may be used to receive radiation simultaneously
passing through the patient. The information received at the
detector can be reconstructed to form an image. Based on this
information, for example, the location of a target may be
identified, and the delivery of the high-energy radiation may be
adjusted as needed.
[0009] Additional advantages and features of the invention will be
apparent from the remainder of this document in conjunction with
the associated drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
[0010] FIG. 1A depicts an example radiation therapy system with
which exemplary real-time tomosynthesis devices can be
implemented.
[0011] FIG. 1B depicts another radiation therapy system with which
exemplary real-time tomosynthesis devices can be implemented.
[0012] FIG. 2A depicts the geometry corresponding to the use of an
x-ray source for fluoroscopy.
[0013] FIG. 2B is a sample detector image resulting from the
radiation beam from the x-ray source of FIG. 2A.
[0014] FIG. 3A shows an exemplary configuration for a tomosynthesis
system that uses multiple imaging sources to intersect at and
illuminate a field of view.
[0015] FIG. 3B depicts an exemplary collimator, viewed from
isocenter, that may be used with the arrangement of FIG. 3A.
[0016] FIG. 4 depicts an exemplary tomosynthesis system that uses
an imager/detector to receive imaging and treatment radiation beams
in different regions/sectors.
[0017] FIG. 5A depicts an exemplary tomosynthesis configuration in
which eight imaging sources are arranged in a ring around a
multileaf collimator (MLC) that can be used to shape a therapeutic
radiation beam.
[0018] FIG. 5B shows an example of eight non-overlapping images
(projections) formed using the imaging sources of FIG. 5A.
[0019] FIG. 5C shows a reconstructed image formed by combining the
eight images of FIG. 5B.
[0020] FIG. 6 illustrates contrast of a tumor as a function of the
number of x-ray tubes (imaging sources) used. The field of view is
5 cm.
[0021] FIG. 7 illustrates real-time tomosynthesis used to track a
moving lung tumor. Columns denote different time points of the
tumor motion. The top row corresponds with fluoroscopy, the middle
row corresponds with real-time tomosynthesis, and the bottom row
corresponds with source 4DCT datasets, 1 cm thick slices. The field
of view is 5 cm.
[0022] FIG. 8 illustrates the impact of tube power on real-time
tomosynthesis. The top row shows AP (anteroposterior) projection,
while the bottom row shows the lateral projection.
[0023] FIG. 9 provides, for comparison, a baseline CT image of a
possible lung tumor, a simulated projection radiograph of the same
lung tumor, and a simulated parallel tomosynthesis image.
[0024] FIG. 10 provides a process of using an exemplary
tomosynthesis system to aid in targeting of a treatment beam in
radiation therapy.
DETAILED DESCRIPTION OF THE PRESENT DISCLOSURE
[0025] FIGS. 1A and 1B depict radiation therapy systems which may
be used in conjunction with exemplary implementations of the
present invention. Referring to FIG. 1A, an example of an
image-guided radiation therapy ("IGRT") system 100 includes a
therapeutic (treatment) x-ray source 102 and a diagnostic (imaging)
x-ray source 104, both of which are housed at an end of a first
rotatable gantry 106 that rotates about a pivot axis 108. The first
rotatable gantry 106 allows the therapeutic x-ray source 102 and
the diagnostic x-ray source 104 to be aligned in a desired manner
with respect to a target volume 110 in a subject 112 positioned on
a patient table 114. A second rotatable gantry 116 is rotatably
attached to the first rotatable gantry 106 such that the second
rotatable gantry 116 is able to rotate about the pivot axis
108.
[0026] Positioned at one end of the second rotatable gantry 116 is
an electronic portal imaging device (EPID), such as x-ray
imager/detector 118. The x-ray detector 118 functions as a
diagnostic image device when receiving x-rays from the diagnostic
x-ray source 104, and can also function as a portal image device
when receiving x-rays from the therapeutic x-ray source 102. The
x-ray detector 118 may contain a number of detector elements (e.g.,
an array of detector elements) that together sense the projected
x-rays that pass through the subject 112. Each detector element
produces an electrical signal that represents the intensity of an
x-ray beam impinging on that detector element and, hence, the
attenuation of the beam as it passes through the subject 112.
[0027] The second rotatable gantry 116 can also include an
articulating end that can pivot about one or more points. The
pivoting motion provided by such points allows the x-ray detector
118 to be moved within one or more dimensions, such as within a
two-dimensional plane. In other configurations, the x-ray detector
118 can be maintained in a fixed position on the second rotatable
gantry 116.
[0028] A control mechanism 120 controls the rotation of the first
rotatable gantry 106 and the second rotatable gantry 116, as well
as operation of the therapeutic x-ray source 102 and the diagnostic
x-ray source 104. The IGRT system 100 includes an operator
workstation 122, which may include a processor 124 that receives
commands and scanning parameters from an operator via an input 126
or from a memory or other suitable storage medium. The input may be
a keyboard, a mouse, a touch screen, or other suitable input
mechanism. An associated display 128 allows the operator to observe
data from the computer 122, including images of the subject 112
that may be used to review or modify the treatment plan, and to
position the subject 112 by way of appropriately adjusting the
position of the patient table 114. The operator supplied commands
and parameters may also be used by the computer 120 to provide
control signals and information to the control mechanism 120.
[0029] The therapeutic x-ray source 102 is controlled by an x-ray
controller 130 that forms a part of the control mechanism 120 and
which provides power and timing signals to the therapeutic x-ray
source 102. The x-ray controller 130 also provides power and timing
signals to the diagnostic x-ray source 104. In some configurations,
the x-ray controller 130 can include two independent controllers
for controlling the therapeutic x-ray source 102 and the diagnostic
x-ray source 104, and in other configurations a single controller
can control both x-ray sources.
[0030] The therapeutic x-ray source 102 produces a radiation beam
132, or "field," in response to control signals received from the
x-ray controller 130. The diagnostic x-ray source 104 projects a
cone-beam of x-rays toward the x-ray detector 118. A gantry
controller 134, which forms a part of the control mechanism 120,
provides control signals to the first rotatable gantry 106 to
control the rotational speed and position of the first rotatable
gantry 106. In response to such control signals, the first
rotatable gantry 106 is moved to change position and the gantry
angle, .theta..sub.i, of the therapeutic x-ray source 102 and the
diagnostic x-ray source 104. The gantry controller 134 connects
with the operator workstation 122 so that the first rotatable
gantry 106 may be rotated under computer control, and also to
provide the operator workstation 122 with signals indicating the
gantry angle, .theta..sub.i, to assist in that control. The gantry
controller 134 also provides control signals to the second
rotatable gantry 116 to change the position and the gantry angle,
.theta..sub.i, of the x-ray detector 118.
[0031] The position of the patient table 114 may also be adjusted
to change the position of the target volume 110 with respect to the
therapeutic x-ray source 102, the diagnostic x-ray source 104, and
the x-ray detector 118 by way of a table motion controller 136,
which is in communication with the operator workstation 122.
[0032] A data acquisition system ("DAS") 138 samples data from the
x-ray detector 118. In some configurations, the data sampled from
the x-ray detector 118 is analog data and the DAS 138 converts the
data to digital signals for subsequent processing. In other
configurations, the data sampled from the x-ray detector 118 is
digital data. The operator workstation 122, or a separate image
reconstructor, receives x-ray data from the DAS 138 and performs
image reconstruction. The reconstructed images can be stored in a
mass storage device, or can be displayed on the display 128 of the
operator workstation 122.
[0033] In an alternative configuration, shown in FIG. 1B, the IGRT
system 100 includes a single rotatable gantry 150 that is movable
to different positions for purposes of administering radiotherapy
and for x-ray image acquisition of a subject 112, as described
above.
[0034] The gantry 150 includes a ring 152, a first arm 154
extending from the ring 152 and to which the therapeutic x-ray
source 102 is coupled, a second arm 156 extending from the ring 152
and to which a portal imager 158 is coupled, a third arm 160
extending from the ring 152 and to which the diagnostic x-ray
source 104 is coupled, and a fourth arm 162 extending from the ring
152 and to which the x-ray detector 118 is coupled. The ring 152 is
rotatable about an axis, such as the pivot axis 108 shown in FIG.
1A, to move the arms 154, 156, 160, 162 (and therefore the
therapeutic x-ray source 102, portal imager 158, diagnostic x-ray
source 104, and x-ray detector 118) about the axis.
[0035] The ring 152 can be driven in any conventional manner,
including without limitation by a sun gear about which one or more
planet gears can be driven, by a ring gear driven by one or more
pinions or worm gears, or by a prime mover directly or indirectly
coupled to an axle upon which a frame is mounted. The motion of the
ring 152 can be controlled by the gantry controller 134.
[0036] The therapeutic x-ray source 102 can include a linear
accelerator (i.e., a "linac") that produces a high-intensity x-ray
beam that exits the therapeutic x-ray source 102 toward the target
volume 110 in the subject 112. In other embodiments, the
therapeutic x-ray source 102 can be replaced with other radiation
therapy devices, such as any device capable of emitting electrons,
gamma rays, and other types of radiation toward the target volume
110 in the subject 112. A variety of radiation-emitting devices
capable of emitting a number of different types of radiation and
adapted for radiotherapy exist and can be implemented in lieu of
the therapeutic x-ray source 102.
[0037] By rotating the ring 152 the arm 154 also rotates, thereby
rotating the therapeutic x-ray source 102 through a range of
different positions about the target volume 110. This adjustment
enables a user to change the trajectory of the radiation beam
exiting from the therapeutic x-ray source 102, thereby enabling the
user to direct the beam to different desired locations in or on the
subject 112. The ring 152 can be rotatable through any range
permitting this beam control. For example, the ring 152 can rotate
through a range of 360 or more degrees in order to move the
therapeutic x-ray source 102 through the same range, although
smaller ranges of movement are also possible.
[0038] As noted above, a portal imager 158 is also coupled to the
ring 152 via the second arm 156. The portal imager 158 can receive
at least some of the radiation from the therapeutic x-ray source
102 in order to generate images of the subject 112. Any
conventional portal imager 158 can be used for this purpose, such
as radiographic film, flat-panel and other types of electronic
portal imagers, and other conventional x-ray imaging devices.
[0039] The portal imager 158 is located opposite the therapeutic
x-ray source 102 across the target volume 110, and can be oriented
to receive radiation emitted from the therapeutic x-ray source 102
as described above. To this end, the portal imager 158 is located
on the second arm 156 extending from the ring 152 at a location
opposite the first arm 154.
[0040] By rotating the ring 152, the arm 156 supporting the portal
imager 158 can also rotate, thereby rotating the portal imager 158
with the therapeutic x-ray source 102 through a range of different
positions about the target volume 110. In this manner, the portal
imager 158 can acquire patient images in the different positions of
the therapeutic x-ray source 102. Although a portal imager 158 is
illustrated in the IGRT system 110 shown in FIG. 2, in alternative
configurations the IGRT system 110 has no portal imager 158.
[0041] As described above, the IGRT system 100 shown in FIG. 1B
also includes a diagnostic x-ray source 104 and an x-ray detector
118 coupled to the gantry 150. The diagnostic x-ray source 104 is
coupled to an arm 160 extending from the ring 152, and the x-ray
detector 118 is coupled to an arm 162 extending from the ring
152.
[0042] By rotating the ring 152, the arms 160, 162 also rotate,
thereby rotating the diagnostic x-ray source 105 and the x-ray
detector 118 through respective ranges of positions about the
target volume 110. This adjustment enables a user to move the
diagnostic x-ray source 104 and x-ray detector 118 in order to
acquire images of the subject 112 taken at different perspectives.
The ring 152 can be rotatable through any range permitting such
control. For example, the ring 152 can rotate through a range of
360 or more degrees in order to move the diagnostic x-ray source
104 and x-ray detector 118 through the same range, although smaller
ranges of motion are also possible.
[0043] The diagnostic x-ray source 104 and x-ray detector 118 are
located across the target volume 110 in circumferential positions
spaced between the therapeutic x-ray source 102 and the portal
imager 158. The diagnostic x-ray source 104 and x-ray detector 118
can be equally or unequally circumferentially spaced between the
therapeutic x-ray source 102 and the portal imager 158. In the
latter configuration, the diagnostic x-ray source 104 and x-ray
detector 118 are located adjacent the therapeutic x-ray source 102
and the portal imager 158, respectively, which can provide
increased access to the subject 112 at one or more circumferential
positions. The exemplary tomosynthesis systems and methods
disclosed herein can be implemented by, for example, integrating
multiple imaging sources with system 100, as further discussed
below. This allows for the repurposing of detector 118 (as well as
the high-voltage power supply for the diagnostic x-ray source 104),
as further discussed below.
[0044] Referring to FIGS. 2A and 3A, the general geometric
principle of exemplary real-time tomosynthesis systems is
illustrated. FIG. 2A depicts conventional fluoroscopy performed
with a source 202, which emits a wide x-ray beam 204 directed
towards detector 206. With the exemplary tomosynthesis system 300
depicted in FIG. 3A, real-time tomosynthesis can be performed by
simultaneous illumination of a field of view (FOV) 302 from
multiple imaging sources 304, 306, 308, which emit imaging beams
314, 316, 318, respectively. A collimator 350 (shown in FIG. 3B)
restricts the radiation beams 314, 316, 318 from each imaging
source so as to intersect at the FOV 222. The exemplary collimator
350 includes multiple exit windows 352 for radiation beams. The
radiation emanating from the imaging sources 304, 306, 308 may be
collimated to, for example, 5 cm by 5 cm at isocenter, and then
these rays diverge and arrive at the detector 320. These
projections from the imaging beams 314, 316, 318 arrive on
different sections of the detector 320. FIG. 2B shows the detector
image for the fluoroscopy arrangement depicted in FIG. 2A, with a
tumor 210 near the center of the image. Fluoroscopy provides poor
tumor contrast with a large field of view (e.g., a 40 cm field of
view). Compared to the fluoroscopy arrangement of FIG. 2A, image
contrast is enhanced with tomosynthesis system 300 (see FIG. 5C,
discussed below). While the FOV 222 with real-time tomosynthesis
may be smaller, the frame rate of fluoroscopy is maintained. It is
noted that the field of view may be enlarged if desired by using
multiple detectors.
[0045] Referring to FIG. 4, an exemplary tomosynthesis system 400
may include a treatment source 402, such as a megavolt (MV) x-ray
source, for treating targeted cells (such as tumor cells). The
treatment source 402 (which may be x-ray source 102 in the systems
depicted in FIGS. 1A and 1B) emits a treatment beam 404 (solid
line) at a tumor 406 in the lung 408 of patient 410. As shown, the
chest of patient 410 is to the left, and the spine of the patient
410 is to the right. That is, from the perspective illustrated, the
treatment beam 404 is directed in a sagittal plane with respect to
patient 410.
[0046] System 400 also includes multiple imaging sources 412, 414,
which may be kilovolt (kV) x-ray sources. Such imaging sources may
be, for example, incorporated into the systems of FIGS. 1A and 1B
by arranging the imaging sources about the therapeutic source 102
(for example, four sources in a ring separated from each other by
90 degrees). Imaging sources 412, 414 may emit imaging beams 416,
418 (dotted lines), respectively, directed such that they
overlap/intersect at a region of interest/field of view that
includes the tumor 406 being targeted by the treatment source 402.
The system further includes collimators 420, 422 for collimating
imaging beams 416, 418, respectively.
[0047] In FIG. 4, the beams emitted by the various sources are
received at a single detector 430 (such as detector 118 in the
systems depicted in FIGS. 1A and 1B), although in alternative
implementations, multiple detectors may be used. In the
configuration shown, the detector 430 may be divided ("sectored")
into regions/sectors that are dedicated to beams from specified
sources, such that no two beams would overlap in the areas of the
detector that would receive the beams. Non-overlapping regions (of,
for example, a flat-panel detector) may have circular, rectangular,
or any other shape desired. Partially overlapping regions could
also be used to expand the field of view, at the cost of
reconstruction complexity and possible artifacts. Treatment beam
404 could thus be received in sector 434, and imaging beams 416,
418 could be received in sectors 436, 438, respectively. Sector 434
may be located, for example, at a central area of detector 430, and
sectors 436 and 438 may have areas arranged around the central
area, such as near the periphery of detector 430 (see FIG. 5B,
discussed below).
[0048] It is noted that any number of imaging sources deemed
suitable may be used, but preferably three or more sources are used
in various implementations. In the exemplary configuration depicted
in FIG. 5A, an exemplary treatment head 500 includes an array of
eight imaging sources (x-ray tubes) 502 arranged in a ring (around
a treatment source, not pictured). The treatment head 500 also
includes a multileaf collimator 504. The x-ray tubes 502 surround
the source of high-energy treatment radiation and are energized
simultaneously or in quick succession, within the time of a single
detector readout. The multileaf collimator 504 controls the
distribution of the high-energy treatment radiation. The number of
sources may vary, and can be based on, for example: cost
considerations (i.e., the cost of additional imaging sources and
installation/assembly thereof); space considerations (i.e., the
number of imaging sources that would fit in the radiation therapy
system); field of view considerations (i.e., additional projections
could crowd the detector and necessitate a reduction in field of
view); the safety of emitting additional x-rays beams at a patient,
as balanced against the benefit of better imaging and thus enhanced
compensation for patient motion via more accurate aiming of the
treatment source (such that damage to targeted tissue is enhanced
and/or collateral damage to healthy tissue is decreased); etc.
[0049] The number of imaging sources 502 can also be based on the
size of the detector (or detectors if more than one is used). FIG.
SB is a simulated image of a detector being illuminated (receiving
data) by parallel tomosynthesis. Each square sub-image corresponds
to a different x-ray source. Specifically, the eight beams from
imaging sources 502 of FIG. 5A may impinge on detector 520 in
different sectors/areas of the detector 520. The larger the
detector 520, the more non-overlapping beams may be simultaneously
received. The central region of the detector remains empty, and
could be used for verification of the MV beam (i.e., therapeutic or
treatment beam). Tumor images 522 are detected at different angles,
and are seen in different sections of the detector. Some of the
projections show a clearer view of the tumor than do others,
however, all of them are obscured to some degree by the overlying
anatomy. They can be combined by tomosynthesis to achieve better
visualization of a tumor, as further discussed below.
[0050] It is noted that the x-ray sources can remain energized
continuously (and detector readouts acquired at the frame rate of
the detector), but this could expose the patient to unnecessary
amounts of radiation. The energizations could instead be limited to
select frames and synchronized to the readout cycle of the
detector, so that the sources are energized and de-energized in
conjunction with the beginning or end of the detector readout. To
reduce scatter from the treatment beam, which may be deleterious to
image quality, it may be necessary to de-energize the treatment
beam during these frames. It is also noted that not all the x-ray
sources need to be energized simultaneously. Certain multi-spot
source architectures may require that only one or a small number of
sources be activated simultaneously. In some cases, the dwell time
per source is substantially shorter than the readout time of the
detector; in these cases, the energizations may be sequential but
be simultaneous from the viewpoint of the detector. In other cases,
groups of sources may be energized at a time. For example, 16
sources could be divided into two groups of 8 sources. In the first
frame, one group of sources may be energized, and in the second
frame, a second group of sources may be energized. This can be done
to increase the field of view on the detector.
[0051] Because power is distributed among many x-ray sources, the
thermal loading on any individual source can be reduced. Smaller
x-ray sources can therefore be used, similar to those used in
dental x-ray imaging. Such sources tend to be more compact (such as
7 cm in length by 3.5 cm in diameter) and less expensive than other
sources, such as cone beam CT sources. They use stationary instead
of rotating anodes. The ring of imaging sources could be placed
next to the multileaf collimator, on the side facing the patient.
The voltage and current demands on the tube filaments is also much
smaller, so a single high voltage generator capable of powering a
CT x-ray tube can be repurposed with a switch to instead power, for
example, ten smaller dental x-ray tubes. The EPID can likewise be
repurposed for tomosynthesis, imaging simultaneously with
treatment. The center region of the EPID would be unusable, as it
would receive a large flux from the MV photons (of the treatment
source). For a collimated SABR treatment, however, the outer
regions would be blank and they could receive kV photons (from
imaging sources).
[0052] To illustrate the viability of real-time tomosynthesis,
images have been simulated using publicly available, online CT
datasets, including a 4DCT dataset that included images of the
tumor in different phases of respiratory motion. In both datasets,
the tumor was about 2 cm in diameter. Forward projection was
performed to simulate the data acquisition process. Monoenergetic
photons were cast through the patient, assuming the attenuation of
water was 0.2 cm.sup.-1 and that attenuation was linear with CT
number. The patient was shifted so that the region of interest was
at isocenter. Relevant geometric parameters in simulations are
listed in Table 1:
TABLE-US-00001 TABLE 1 kV source-isocenter distance 40 cm
Isocenter-EPID distance 40 cm Multileaf collimator diameter 25 cm
Source ring diameter 30 cm (focal spot location) Number of x-ray
sources 8, 16, or 32 EPID detector dimensions 41 cm by 41 cm Field
of view 5 cm EPID dimensions 1024 .times. 1024 pixels Pixel pitch
400 um X-ray tube size 7 cm (length) by 3.5 cm (diameter)
[0053] Reconstruction from the detector can be done using, for
example, shift-and-add, or can be done with filtered
backprojection, or with iterative methods. If reconstruction is
performed with shift-and-add, a high-pass filter can be used to
accentuate detail in the plane of focus. Here, after the detector
image was generated, reconstruction was performed by shift-and-add,
which is a relatively computationally efficient process. FIG. 5B
shows individual images of projection radiography from different
angles, and 5C shows an example of these detector images summed
together for reconstruction, with improved contrast. In other
images presented in this work, shift-and-add is followed by a
high-pass filter, which is equivalent to a filtered
backprojection-type reconstruction. Iterative reconstruction may
also be used, but iterative reconstruction at 30 frames per second
may be challenging to implement. FIG. 2B provides a radiograph
showing a tumor 210, partially obstructed by overlying anatomy.
FIG. 5C provides a zoomed-in image from real-time tomosynthesis
(with tumor 530 near the center), produced by shifting and adding
the sub-images shown in FIG. 5B prior to any high-pass filter. The
tumor is better resolved in FIG. 5C than the tumor in FIG. 2B.
[0054] The quality of tomosynthesis depends in part on the number
of views used to generate the image. By comparing different
configurations, different tradeoffs in image quality and complexity
can be observed. One potential configuration is to array the
imaging x-ray sources in a simple ring. With the assumptions in
Table 1, there is enough room on the EPID for approximately eight
images without any overlap. The ring is wide enough so that the
ring does not block the exit path from the multileaf collimator.
The x-ray sources are small enough so that there is no crowding in
the source ring. As a single dental x-ray tube requires half the
filament current and a quarter of the filament voltage of a more
powerful CT x-ray tube, the eight small tubes can be powered using
the same x-ray generator already used for the onboard CBCT (cone
beam computed tomography) scanner.
[0055] To achieve a greater number of views, a ring configuration
with 16 sources may be used in various implementations. To prevent
overcrowding of the detector (or overlapping images), each source
may be energized only in every other view. Assuming half are
energized in each view, the imaging time in this arrangement would
be half the frame rate of the EPID. Such a configuration may
include a high-voltage switch working in synchrony with the
detector, and the switch may add a modest amount of hardware
complexity in certain implementations. Further increasing the view
count (for example, 24 or 32 sources) is also possible.
Significantly more than 24 sources may lead to overcrowding of the
source ring, or collision of the tubes. This could be mitigated by
arranging the tubes in layers, but this would tend to increase the
thickness of the source ring.
[0056] Extended image guidance for radiation therapy could impart
additional dose to the patient. Even if this dose is much smaller
than the therapeutic radiation, unnecessary dose should be avoided
when possible. In examining the effect of different power levels on
the reconstructed images, it can be observed that the absorbed dose
is proportional to the power on each source. It is also observed
that x-ray tube cooling requirements may limit the total power
available to the system. In simulations, Poisson noise was injected
into the detected images according to the number of photons
reaching the detector.
[0057] It was assumed that the fluence arriving on the detector was
5.times.10.sup.5 photons/mAs-mm.sup.2 at a distance 80 cm from the
focal spot. The conversion factor varies depending on added
filtration and tube design, but this is a reasonable estimate for a
tube operating at 80 kVp (kilovoltage peak). It was assumed that
the pulse duration in all cases was 25 ms and that the tube power
was set to either 0.5 mA or 2 mA. Hence, for the assumed pixel
pitch of 400 um, the number of photons reaching the detector in the
absence of patient attenuation is either 1000 or 4000 per frame.
Images from both the AP and lateral direction were examined. In the
lateral direction, patient attenuation is much higher and hence
higher doses could be necessary to ensure adequate
visualization.
[0058] FIG. 6 shows a comparison between real-time tomosynthesis
and the source CT dataset as well as fluoroscopy. Specifically, the
contrast of a tumor as a function of the number of x-ray tubes used
is shown. From left to right, FIG. 6 shows: 1 cm CT slice (602); 32
views (604); 16 views (606); 8 views (608); and 1 view (i.e.,
fluoroscopy) (610). While the tumor is virtually invisible on
fluoroscopy (610), its location can be seen with tomosynthesis
(604, 606, 608). The image quality is better with a greater number
of views, but the frame rate is reduced for images shown. Assuming
a detector frame rate of 30 fps, using 8, 16, or 32 views would
yield a system frame rate of 30, 15, or 7.5 fps, respectively. A
reduction of frame rate may be unacceptable in case of excessive
motion. It is noted that the high-pass filter used to isolate the
plane of interest has yielded some artifacts in the images, such as
a dark region around the lesion.
[0059] FIG. 7 shows an example of tracking the same lung tumor over
time. Different phases of a 4DCT series were used to generate the
simulations. Columns denote different time points of the tumor
motion. The top row (702) corresponds with fluoroscopy, the middle
row (704) corresponds with real-time tomosynthesis involving 16
x-ray sources, and the bottom row (706) corresponds with source
4DCT datasets. The tumor can be seen with fluoroscopy (702), but
its visibility is limited compared to real-time tomosynthesis
(704).
[0060] FIG. 8 shows images of real-time tomosynthesis corresponding
with the use of 16 views (15 fps), from both the lateral and AP
(anteroposterior) directions, shown with two different power levels
as well as the noiseless case. The top row (802) shows AP
projection, while the bottom row (804) shows the lateral
projection. The left column (810) is the noiseless case (i.e.,
without any added noise). The middle column (812) includes noise
corresponding to each tube operating at 2 mA, or 60 W average power
draw. The right column (814) includes noise corresponding to each
tube operating at four times less power. In the higher power
setting, each of the 16 sources was energized for 25 ms at 2 mA and
80 kVp. In the lower power setting, each source was set to 0.5 mA.
At any point in time, the total power draw of the source ring will
be 8 times higher because 8 sources are active. However, because
each source is tightly collimated, the dose imparted to the patient
will be relatively small.
[0061] At the lower power setting, because the duty cycle on each
tube is about 40%, the average power draw for any individual tube
is 15 W. The tumor is highly visible even at 15 W of power per
tube. It is noted that in video mode, the observer may be able to
tolerate higher levels of image noise than are present in still
images.
[0062] For comparison, FIG. 9 provides a baseline CT image (902) of
a possible lung tumor 910, a simulated projection radiograph (904)
of the same lung tumor 910, and a simulated parallel tomosynthesis
image (906) showing the tumor 910. 16 sources are used, and a
high-pass filter is also included in the reconstruction to
accentuate detail in the plane of focus. The overlying bone ribcage
is largely eliminated.
[0063] An exemplary tomosynthesis process 1000 is depicted in FIG.
10. To locate a target, a region of interest suspected of having
the target can be simultaneously illuminated with multiple imaging
sources (1005), such as kV x-rays that are situated about a
treatment source, such as a MV x-ray (capable of emitting
high-energy treatment radiation that is, for example, above 200
keV). The beams from the multiple imaging sources can be received
at one or more imagers/detectors (1010), such as a detector that is
sectored so that area that is not otherwise used by a treatment
source is sectionalized for the imaging sources. A readout from the
detector containing the multiple images/projections from the
imaging sources can be acquired, and images combined to form a
reconstructed image (1015) using, for example, a shift-and-add
procedure. It is noted that if the detector functions at 30 fps,
the detector could be set to acquire 30 readouts per second. The
reconstructed image could be analyzed to locate the tumor; this
could be performed automatically using, for example, computer
vision/image recognition, or via recognition by a user (such as a
radiologist) and identification of the tumor (or a portion of the
image showing the tumor) via a user input device (such as a mouse,
touch screen, etc.).
[0064] The direction/aim of the treatment source, such as a MV
x-ray, can be set or adjusted as needed (1020) so as to align the
treatment beam to maximize overlap with the target (e.g., a tumor)
and minimize overlap with surroundings (such as normal or otherwise
not-targeted tissue). The treatment/therapeutic source can then be
used to apply radiation to the target (1025). Either simultaneously
with the application of radiation, or thereafter, the region of
interest may be irradiated again using the imaging sources (1005),
and the process iteratively repeated (with location of tumor and
adjustment of treatment source) for the duration of a treatment
regimen. The operation of the imaging sources can be controlled by
a controller (such as x-ray controller 130 in FIG. 1A) that
controls the treatment source, or in other implementations, using a
separate controller in communication therewith. A controller may
include, for example, a processor and memory with instructions
executable by the processor to perform the above functions. The
controller may be implemented using hardware, software, or a
combination thereof.
[0065] It is noted that the location and size of the field of view
can be set so that it encompasses the potential locations of the
moving target. For example, if during normal breathing, the tumor
may rise and fall across a distance of 3 cm, then the field of view
may span a height of 3 cm. Although the imaging sources are at
lower energies than the treatment sources, unnecessary exposure to
x-rays should be avoided. Consequently, the field of view can be
configured, in certain implementations, to be adjustable as
warranted. For example, if a patient's breathing changes such that
the tumor no longer rises and falls across a span of 3 cm but
rather 2 cm (e.g., if the patient relaxes and starts taking
shallower breaths), then the field of view can be shrunk (for
subsequent images) to reduce the volume receiving imaging x-ray
beams. Similarly, because it is possible for a patient to move so
that the tumor reaches an "outlier" position (e.g., if the patient
has a muscle spasm or is bumped), the system may be configured to
resist expanding the field of view unless a target moves beyond the
existing field of view a predetermined number of times or for a
predetermined length of time. In other words, in exemplary
configurations, the system need not "look" where the target is not
expected to be (or is unlikely to be in the future), because
"looking" in such places unnecessarily exposes the patient to more
radiation.
[0066] Applied to radiation therapy, components that already exist
on the linac may be repurposed, including the high-voltage power
generator and the EPID. The additional hardware used, if not
already included, is an imaging source ring (including, for
example, 8 to 16 dental x-ray tubes in a single housing) and a
collimator. The source ring and collimator could be placed directly
on top of the existing multileaf collimator, and could be
engineered to add, for example, 4 cm of thickness to the multileaf
collimator. Moreover, a high speed kV/MV detector (or EPID)--i.e.,
a detector that can detect both kV photons and MV photons--may be
used.
[0067] While imaging from the beam's eye view is desirable, in some
cases it may be more convenient to image from other directions. In
some cases, an auxiliary detector and source are part of the
treatment system. This could include the cone-beam CT imager, or it
could be a secondary system permanently affixed to the room. In
these cases, a source ring and collimator could be installed in
proximity to the original source. Depending on the need, the system
could either select the original source for a full, field of view
image, or it could select the source ring to produce tomosynthesis
in a reduced field of view.
[0068] The above approach can be applied to many different
treatment applications. In cases of motion management by limitation
(in particular, patient-initiated breath hold), exemplary
tomosynthesis systems could serve as a verification device that
could promptly halt treatment if the tumor is out of the ITV. In
motion-inclusive or respiratory-gated treatments, the ability to
follow the tumor motion track in real-time could potentially allow
for a decrease in the ITV and planning target volume (PTV) margins.
Clinically, this would reduce the volume of normal lung treated,
thus possibly reducing the risk of radiation pneumonitis. In the
case of central tumors, this approach may allow tumors previously
deemed "too close" to central airways or major vessels to be more
safely treated. As compared with external surrogates, direct
visualization may enable greater accuracy of tumor tracking.
Real-time tomosynthesis is also likely more comfortable for
patients than some existing options for motion management, such as
abdominal compression or active breathing control.
[0069] Simulated images of real-time tomosynthesis show good image
quality and contrast of the lung tumors, with a noticeable
improvement over fluoroscopy. The lung is a particularly useful
case because lung nodules have strong contrast against the
background. Whereas fluoroscopy might not be able to identify
certain central lung tumors due to poor contrast, the real-time
tomosynthesis approach discussed can be utilized in such cases.
Because the contrast in the lung is so high, and because several
tubes are used simultaneously, the power demanded from each tube is
relatively low, enabling the use of compact, stationary anode x-ray
tubes.
[0070] Compared to conventional tomosynthesis, the exemplary
systems for real-time tomosynthesis disclosed involves acquisition
of images much more quickly, but the field of view may be
relatively smaller with one sectored detector. This makes it
well-suited for lung SABR treatments, where the tumors treated tend
to be relatively small (i.e., under 5 cm). Because the field of
view is smaller, the dose imparted by real-time tomosynthesis may
be relatively small compared to fluoroscopy. For example, while
eight tubes may be activated simultaneously, for any given power
level, the dose-area-product of a single tube will be 16 times less
because only a small portion of the detector will be irradiated. In
addition, the example studied showed good visualization of the
tumor even at less than 1 mA. The EPID sometimes includes a thin
sheet of copper to improve detection efficiency of MV photons. This
copper was not modeled in simulations discussed here but could
impede the detection of kV photons. This effect could be reduced or
eliminated if the copper sheet were retracted. For 60 keV photons,
1 mm of copper would absorb approximately 75% of the incident
photons. If the copper were not retracted, the dose and power
requirements of real-time tomosynthesis would likely be higher but
may still be acceptable.
[0071] In alternative implementations, the real-time tomosynthesis
approach is not limited to radiation therapy. The approach is
applicable to, for example, interventional radiology. In this
context, a moving field of view may be desired to track an
instrument. This can be achieved by attaching the collimator to a
three-axis motor and using it to follow the volume being imaged. A
mechanized collimator could also be used for the treatment of
tumors that are not at isocenter. In some applications, the use of
multiple detectors or reduced magnification may be necessary to
enlarge the field of view. In other implementations, the above
approach may be applied to ablation for arrhythmia, something that
requires careful image guidance due to rapid cardiac motion. Also,
the spine is a common target for SABR and demands high accuracy,
and as a high-contrast object, it could be tracked using
tomosynthesis. Outside of radiation therapy, minimally invasive
interventions may be guided using real-time tomosynthesis. For
example, exemplary tomosynthesis devices could be used to obtain
diagnostic images in the thorax with a minimum amount of
motion.
[0072] Exemplary systems and methods for tomosynthesis are capable
of imaging a small field of view at the full frame rate of the
detector. In various implementations, the approach uses parallel
acquisition of multiple frames by simultaneously illuminating the
field of view with multiple sources. EPIDs, or kV/MV detectors,
used in linear accelerators can operate at up to 30 fps. For a
modest number of views, the imaging time with tomosynthesis will
therefore be on the order of a second, which may be insufficient to
resolve respiratory motion. In light of this constraint, exemplary
implementations may use a single detector readout to measure
multiple views simultaneously. In the lung, for example, SABR is
usually applied only for tumors with diameters measuring 5 cm or
less. The EPID may measure, for example, 41 cm by 41 cm in area.
The EPID thus has the surface area to resolve multiple
non-overlapping projections of the small tumor.
[0073] In other implementations, a device for parallel
tomosynthesis includes a plurality of x-ray sources that are
energized simultaneously or in rapid succession, causing an object
being scanned to be illuminated from multiple viewpoints. The
resulting detector will include the x-ray projections (or shadow
images) from all the sources, where the projections are located at
different parts of the detector image (because the detector is
sectored). One detector readout is thus sufficient to acquire
multiple views that can be used to reconstruct an entire
tomosynthesis image. This reduces the number of detector readouts
necessary, and enables tomosynthesis at frame rates sufficient to
capture, for example, respiratory or cardiac motion. It is possible
to build a plurality of x-ray sources that are simultaneously
energized. However, it is also possible to toggle between the
miscellaneous x-ray sources in rapid succession, such that a
plurality of these sources is energized within the time frame of a
single detector readout. In both cases, the advantage of real-time,
parallel tomosynthesis may be realized.
[0074] Exemplary x-ray sources surround a radiation therapy
delivery device that transmits high-energy radiation, often photons
in the megavoltage range, using a mechanism such as a linear
accelerator or a radioisotope. Examples of these devices include
external beam radiation therapy machines or robotic radiosurgery
systems. The x-ray radiation from each source is collimated so that
the images from each view end on different portions of the
detector. Reconstruction would be simpler if the images for each
view are separate and distinct from each other, although some
overlap of the projection images may be tolerated. Because the
x-ray sources are illuminated simultaneously, simpler and lower
power x-ray sources can be used. For example, standard stationary
anode x-ray tubes may also function well.
[0075] In the context of radiation therapy delivery, it may be
advantageous to use the detector ("EPID" in external beam
radiotherapy) already present on some systems to detect the
high-energy beam because this type of detector is able to image
megavoltage photons as well as kilovoltage photons. Especially in
stereotactic radiotherapy or radiosurgery applications, the
high-energy treatment beam may be contained so that only a small
field of view is exposed to therapeutic radiation, while most of
the field of view of the detector is unused. The regions near the
periphery of the detector could be repurposed and detect
kilovoltage diagnostic x-ray photons for exemplary implementations
of the tomosynthesis approach disclosed here. This can be done
simultaneously with radiation treatment, or time for the
tomosynthesis and the time for radiation treatment could be
interlaced.
[0076] In external beam radiation therapy devices in particular,
the introduction of real-time tomosynthesis could have relatively
modest hardware requirements. Many of these machines already have
EPIDs, x-ray generators, and flat panel detectors. A ring of
stationary anode x-ray sources could be wired together on the
treatment arm. These sources may be on the order of an inch in
diameter and would be fairly compact. These sources would be
positioned around the multileaf collimator, and they themselves
would be collimated down to a small region in isocenter. During
treatment, these sources would be periodically pulsed and the EPID
detector would be read out. The central portion of the EPID could
be used for treatment verification, but the outer portions would
contain data for several views and can used for tomosynthesis.
[0077] In radiosurgery devices that do not rigidly fix the patient
down, positioning accuracy is also critical. A new detector arm
could be built and could house a series of smaller kV detectors, or
a single large kV/MV detector that would also perform verification
of the treatment beam.
[0078] Tomosynthesis performed in this manner would capture the
tumor from the perspective of the treatment beam and would thereby
be directly relevant for treatment. Motion of the tumor can be used
to inform the location of radiation treatment. For example, for
systems using multileaf collimators, the leaf positions could be
adjusted to follow the tumor. Alternatively, the treatment itself
could simply be gated. It is noted that if the target moves
parallel with the treatment beam (i.e., forwards or backwards), the
treatment source need not be re-aimed because the target would
remain in the path of the treatment beam. If, however, the target
moves left/right or up/down, then the treatment beam may be shifted
accordingly to better track the target.
[0079] This type of tomosynthesis may have other applications in
diagnostic radiology. It can be used to take a single frame in the
presence of respiratory or cardiac motion with much higher temporal
resolution than architectures which require multiple detector
readouts and possibly motion of the x-ray source.
[0080] In other configurations, several small detectors can be
operated in parallel. These detectors could be spaced farther
apart, increasing the tomographic angle and hence the image quality
in some applications.
[0081] In certain implementations, performing parallel
tomosynthesis as described above may limit the number of views that
can be simultaneously acquired. To improve the image quality, it
may be desirable to group the x-ray sources into a number of
subsets, and to fire each subset at a time. For example, if N
sources were labeled 1, 2, 3 . . . N, then one imaging strategy is
to energize the odd sources in the first detector readout, and the
even sources in the next detector readout. This would increase the
number of views but would require two readouts.
[0082] The present disclosure has described one or more preferred
embodiments, and it should be appreciated that many equivalents,
alternatives, variations, and modifications, aside from those
expressly stated, are possible and within the scope of the
invention.
* * * * *